Review of U.S. Department of Transportation Truck Size and Weight Study - Second Report: Review of USDOT Technical Reports (2015)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

10 2. GENERAL OBSERVATIONS The conclusions below concern the extent to which the technical reports as a whole address the issues identified by Congress and the appropriateness of methods and data that affect all parts of the USDOT study. The recommendations concern information needs and the structure of future truck size and weight policy studies. Responsiveness to the Questions Identified by Congress The USDOT report carefully specifies how the study attempted to address each of the requirements of the congressional study charge (Summary, 15–18) and acknowledges gaps in its ability to estimate impacts of alternative configurations. However, the experience of past truck size and weight studies suggests that the present study could have been more responsive to the needs of Congress and its other audiences. Limitations of the USDOT study that could have been at least partially overcome are identified below. Lack of a Summary Evaluation The USDOT report lacks a consistent and complete quantitative summary of the evaluations of the alternative configuration scenarios. In the absence of such a summary, the report is not fully responsive to the congressional charge to assess the impacts of alternative and exempt vehicles. Moreover, the lack of a summary diminishes the report’s value to its intended audiences. The 1981 USDOT truck size and weight study and past TRB studies contained summaries of all impacts considered of specified changes in size and weight regulations, with most impacts expressed in terms of nationwide dollar-denominated annual costs and benefits (USDOT 1981, V-6; TRB 1990a, 13– 15; TRB 1990b, 180–181). The 2000 USDOT study included a summary table of the percent changes in each cost and benefit, excluding safety, for each regulatory scenario (USDOT 2000, 38) and presented

11 estimates of nationwide dollar impacts in the body of the report. The tabular summary in the 2015 USDOT report (Summary, ES-11–ES-12) presents findings for some, but not all, costs considered, and with such disparity in definitions and metrics that comparison among the scenarios is impractical. Omission of a summary evaluation was a decision of the 2015 USDOT study’s authors. USDOT concluded that “data limitations are so profound that results cannot accurately be extrapolated to predict national impacts.”2 However, projections of the kind contained in the 2015 and earlier studies cannot be characterized simply as accurate or inaccurate. Each estimate possesses some degree of uncertainty, which must be described. None of the past studies used models or data that were more accurate than those of the 2015 USDOT study. The reports of some of those studies included adequate discussion of critical uncertainties and of the role of judgment in estimating impacts, while others did not. The 2002 TRB committee that reviewed all major past truck size and weight studies concluded that, because of the complexity of the highway freight transportation system, “it is not possible to predict the outcomes of regulatory changes with high confidence” (TRB 2002, 3). If highly confident forecasts of impacts were the standard, no regulation would ever be enacted or changed. Changes at the federal and state levels are being considered or enacted today, and government transportation professionals need to provide the best estimates available at any time to aid these deliberations. Notwithstanding the USDOT study’s conclusion that some costs of the regulatory alternatives are impossible to estimate confidently, the report could have provided a framework for understanding all the costs and benefits that would help users to put the costs that are estimated in perspective. A comprehensive list of the categories of costs and benefits, the features of the regulatory change that influence each category (e.g., pavement wear is affected by axle weight limits, bridge costs by gross weight), the expected direction of change, and the cost categories that are likely to be critical to the evaluation all can be identified on the basis of the results of the 2015 USDOT study and past size and weight studies. A useful summary would correct three shortcomings in the USDOT report’s analysis and 2 Letter of Peter M. Rogoff to Hon. Bill Shuster, June 5, 2015.

12 presentation: missing estimates of categories of costs and benefits, inconsistent units of measure of impacts, and inconsistent and sometimes misleading assessments of the uncertainty of estimates. These limitations are examined in the sections below. Missing Costs Categories of costs that were considered in past studies or that are likely to be affected by changes in size and weight regulations but were not estimated in the USDOT study include the following:  Infrastructure costs on roads not part of the Interstate system or National Network;  Bridge deck maintenance and rehabilitation costs;  Expected or likely bridge structural costs [the cost estimates reported are described as an “extreme upper bound,” and the USDOT study concludes that “neither the actual costs nor the lower bound costs are determinate” (Bridge, 58)];  User costs of bridge replacements and retrofits and of bridge weight restrictions;  Aggregate crash and casualty frequency and associated costs;  Crash risk in highway work zones; and  Costs in future years, when freight traffic and total highway traffic will be greater (costs and benefits are unlikely to grow proportionally with traffic volume). In some cases (e.g., crash frequency, bridge decks, local roads), the USDOT report acknowledges that a cost would occur, although it does not estimate the cost. Other important costs are not mentioned in the body of the technical analysis, although estimates of past studies may be cited in the desk scans. The committee recognizes that producing definitive estimates of all of these costs would not have been practical within the constraints of time and resources available for the USDOT study. The critical

13 uncertainties in evaluations of changes in size and weight limits, according to the 2002 TRB committee review of the problem (TRB 2002, 3-4), are safety consequences and bridge costs. Differences in crash rates that depend on vehicle weight or configuration will remain difficult to measure until improvements are made in the systems for regular monitoring of traffic volumes and for recording characteristics of vehicles involved in crashes. An appropriate bridge cost estimate (as described in Chapter 3 below) would require development of new methods and might not have been feasible within the study’s resources. Nevertheless, the USDOT study could have produced order-of-magnitude estimates of the important costs, with confidence intervals derived from the range of estimates produced in the 2015 study as well as past studies and from sensitivity analysis. Moreover, the study could have presented its estimates in a comprehensive, consistent framework defining national costs and benefits. Such a presentation would be useful to Congress and the public in considering proposed changes to regulations. It would show the overall magnitude of potential benefits and costs at stake in regulatory decisions; indicate how impacts would be distributed among shippers, carriers, highway agencies, highway users, and the public; and possibly reveal some actions that promised benefits at low risk. Such a comprehensive summary evaluation also is necessary for identifying priorities for research on measuring impacts and on mitigating the costs of truck transportation. Inconsistent Metrics Inconsistency in the units of measure in which the impacts estimates are presented makes it impossible for the reader to weigh the trade-offs that are inherent to any regulatory decision:  Changes in logistics costs, congestion costs, and enforcement costs are reported in annual dollars, in the analysis base year of 2011 (i.e., 2011 freight volume, highway traffic, and prices are assumed).  Changes in fuel consumption and pollutant emissions are reported in physical units (gallons of fuel and pounds of emission) on an annual basis.

14  Costs of accommodating bridge structures to heavier gross weights are reported as an upper bound of the one-time capital expenditure required for replacement or retrofit of bridges.  Pavement wear costs are presented as a percentage change in life-cycle cost (i.e., the percentage change in the present value of highway agency costs for construction and rehabilitation of pavement in all future years).  Safety impacts are described in a bullet-point list of various qualitative and quantitative observations about crash rates and stability properties of alternative configurations. Use of a single unit of measure, at least for the costs most readily expressed in dollar terms, would have enhanced the usefulness of the USDOT study. The most comprehensive measure would be the present value of future changes in each cost category. Alternatively, impact estimates could be presented in annualized dollars in the analysis year (2011) and in one or more future years to illustrate the effect of traffic growth on costs and benefits. Inconsistent Assessments of Uncertainty The USDOT report provides insufficient quantitative indication of the degree of uncertainty of its impact estimates. Users of the report need indications of the degree of confidence to be placed in each impact estimate. No simple way of defining or expressing uncertainty in such a complex study exists; however, past truck size and weight studies have used several techniques effectively:  Comparison and synthesis of evidence from all available sources, including comparison with estimates from past studies. For example, the TRB committee that evaluated twin trailer trucks based its estimates of crash rates on a critical review and synthesis of 16 past studies. None of the 16 studies individually was conclusive, but the TRB committee concluded that collectively their results

15 narrowed the range of uncertainty about safety impacts (TRB 1986, 320–324). The desk scans prepared for the USDOT study summarize results of past studies, but the USDOT report makes little use of these results, either to supplement the study’s own estimates or as a check on the credibility of the new estimates.  Comparison of estimates derived by alternative independent methods. If two independent methods yield similar estimates, confidence is improved. For example, this approach could have been used to help assess the validity of the sample of bridges used in the bridge structural cost estimates in the USDOT study.  Estimates of statistical uncertainty (e.g., the uncertainty of an estimate derived from information on a sample of the population of interest or based on errors of measurement in the inputs to an impact).  Highlighting of uncertainties that are most important for decision making. The tolerable magnitude of uncertainty of an impact estimate depends on the relative magnitude of the impact. For example, past studies have established that, for changes in regulations involving changes in gross weight limits but no change in axle weight limits, costs of pavement impacts are much smaller than costs of bridge impacts. Therefore, devoting more study resources to improving estimates of pavement impacts would not contribute greatly to the accuracy of the overall assessment of infrastructure impacts of such a regulatory change. An estimate of an upper bound on a cost (as the USDOT study estimated for bridge structural costs) is sufficient if the cost is known to be small compared with other costs. However, for major costs (including bridge structural costs of increases in gross weight limits), an estimate of expected cost with a confidence interval is needed.  Sensitivity analysis. For example, the TRB committee that authored the report on the Turner proposal used sensitivity analyses to show how bridge costs would be affected by alternative assumptions about responses of highway agencies to introduction of heavier trucks and to show how alternative projections of market penetration of new truck configurations would affect highway agency costs (TRB 1990b, 191–194, Tables 7-2 and 7-3), and the TRB Truck Weight Limits committee used

16 sensitivity analysis to show how alternative assumptions about the relationship of truck weight to crash rate would affect the estimates of safety costs of changes in limits (TRB 1990a, 252–253). The USDOT report does not make systematic use of sensitivity analysis. The past USDOT and TRB studies used these techniques to support incisive conclusions about the probable range of effects of changes in size and weight limits, in spite of the gaps in knowledge that have confronted all such studies. Definition of Regulatory Alternative Scenarios The USDOT study estimates costs and benefits for six scenarios. Each scenario is defined in terms of a single vehicle not allowed under present federal law that federal law would allow to operate on the Interstate system and the federal designated National Network (or on a more restricted network in the triple-trailer scenarios). In contrast, the earlier USDOT and TRB studies defined scenarios in terms of changes in regulations (e.g., a change in the gross weight limit). Analyzing vehicles one at a time is insufficient for estimating the effects of regulatory changes that would lead to increased use of more than one configuration type because the impacts of introducing two of the alternative configurations would not equal the sum of the impacts of introducing each one individually. The alternative configuration scenario definitions reflect the point of view expressed in USDOT’s presentation to the committee on the technical reports, which defined the “fundamental truck size and weight policy question” as follows: “Do the estimated ‘positive’ impacts of a particular TSW [truck size and weight] change outweigh the estimated ‘negative’ impacts?”3 The TRB committee that authored Regulation of Weights, Lengths, and Widths of Commercial Motor Vehicles (TRB 2002) argued that this definition of the size and weight policy problem is insufficient and that, instead, a truck size and weight policy study should begin by defining the objectives of size and weight regulations [which the 2002 TRB 3 FHWA presentation to the committee, July 14, 2015.

17 committee defined as “balancing the potential public costs of truck travel against the benefits of lower shipper and carrier costs for freight transportation” (TRB 2002, 41)] and then should search for changes in the regulations and other government policies that would better meet the objectives (TRB 2002, 41–42). For example, if evaluations indicated that a change in limits would allow reductions in shipper costs but that bridge-related costs and uncertainty about safety were obstacles, the size and weight study should next consider whether means of overcoming these obstacles may exist (e.g., by identifying vehicle design requirements to improve safety, cost-effective strategies for mitigating bridge impacts, or fee schemes to fund necessary improvements). The legislative study charge requires USDOT to evaluate “alternative configurations,” but nothing in the charge precludes the approach recommended by the 2002 TRB committee. The charge does not specify the form of the change in federal law (e.g., a change in the federal dimensional limits or an exemption from the limits for a specified configuration) that would lead to introduction of the alternative configurations. In defining the alternative configuration scenarios of its study, USDOT found it necessary to specify various aspects of the regulatory provisions that would govern the alternative configurations, including road networks and axle weight limits. Evaluation of Consequences of Grandfather Exemptions Congress charged USDOT to “evaluate the impacts to the infrastructure in each State that allows a vehicle to operate with size and weight limits that are in excess of the Federal law and regulations, or to operate under a Federal exemption or grandfather right, in comparison to vehicles that do not operate in excess of Federal law and regulations . . ., including—(A) the cost and benefits of the impacts in dollars; (B) the percentage of trucks operating in excess of the Federal size and weight limits; and (C) the ability of each State to recover the cost for the impacts, or the benefits incurred. . . .” [MAP-21, Section 32801(a)(2)]. USDOT was also to “provide data on accident frequency and evaluate factors related to accident risk of vehicles that operate with size and weight limits that are in excess of the Federal law and regulations in

18 each State that allows vehicles to operate with size and weight limits that are in excess of the Federal law and regulations. . . .” [MAP-21, Section 32801(a)(1)]. These directions appear to call, at least, for USDOT to provide an overall assessment of the consequences for infrastructure, safety, and state finances of the grandfather and other exemptions in federal size and weight regulations in the states where the exemptions are in effect. The 1981 and 2000 USDOT truck size and weight studies directly addressed the question of the consequences of the grandfather exemptions by assessing regulatory scenarios in which the exemptions were eliminated (USDOT 1981, II-2; USDOT 2000, II-11). Estimates of costs and benefits in these scenarios in the earlier studies were presented for the nation, rather than at the state level, although the 1981 study discussed state-level industry economic impacts. The 2015 USDOT study does not provide an overall assessment of the consequences of the grandfather exemptions in the states where they apply. It therefore falls short of responding to this part of the congressional charge. Methods and Data The observations in this section concern study methods that affect multiple impact areas. Chapters 3 through 7 below present the committee’s conclusions concerning methods and data used in the USDOT study’s estimates of each of the five impact categories. Limitations of Advanced Models The USDOT study cites its use of improved models and data sets as an advance over the previous USDOT studies (Summary, 31–32). The most current methods are used in the pavement, bridge, and safety analyses. The modal shift and truck travel estimates used traffic information from FHWA’s Freight Analysis Framework, which did not exist at the time of the previous studies. Certainly, it was appropriate for the study to examine whether these new resources could provide new insights into the size and weight

19 problem. However, the data requirements and other features of the new models imposed restrictions on the study. The bridge analysis examined a sample of 490 bridges, whereas the 2000 study analyzed more than one-fourth of all U.S. bridges with a simpler method. In the pavement analysis, the model cannot represent all pavement deterioration mechanisms or load impacts on pavement overlays. The USDOT study might have been able to complete or confirm some of the impact estimates in the report by using older models alongside the new ones. 4 Comparison of results of new and old models also would have been valuable as a test of the benefits of the new models. Many states have more reliable, detailed, and comprehensive data today about infrastructure condition and truck weights and traffic volumes than were available for the 1981 and 2000 USDOT truck size and weight studies, as a result of progress in infrastructure management programs. Taking advantage of improved data, rather than enhanced models, might have been the greatest opportunity for advancing analysis of size and weight policy over the earlier studies. Inaccurate Descriptions of Uncertainties Estimates in several of the impact categories that involve statistical calculations or that include confidence intervals inappropriately characterize uncertainty:  The tests of significance of differences in crash rates between pairs of vehicle types (Safety, Tables 8– 10) are incorrect. The test applied assumes that the values of the denominators in the ratios compared are known with certainty, but in this computation, the values of the denominators [vehicle miles of travel (VMT) for each vehicle type] are uncertain to an unknown degree.  In the presentation of the results of the model relating crash frequency to traffic volume (Safety, 29– 30, Figures 2 and 3), the extrapolation of the relationship derived by regression far beyond the range 4 The bridge analysis used an alternative method of structural evaluation for bridge types that the preferred method could not represent (Bridge, 13–14).

20 of the observations is unjustifiable because the extrapolated crash frequencies have extremely large uncertainties.  The table presenting estimates of the change in life-cycle pavement costs (Pavement, 28, Table 12; Summary, 58–59, Table 9) shows ranges of values, as if indicating the uncertainty of the estimates. However (as the text explains), the high and low values shown are simply the estimates for two assumed values for the discount rate in the life-cycle analysis. The report does not analyze the effect of any other source of uncertainty.  In the bridge analysis, the report explains that the sample size of 490 bridges selected for structural analysis was determined according to a calculation of the sample size required to achieve a desired confidence level in estimates derived from the sample (Bridge, 16). However, because the sample was not randomly selected, this calculation is inapplicable. Truck Travel Estimates from Weigh-in-Motion Data The baseline estimates of annual VMT by each truck configuration used throughout the study (for estimates of bridge and pavement costs and alternative configuration crash rates) are derived from data collected by the states from weigh-in-motion (WIM) devices. The committee’s 2014 report noted the difficulties encountered by others in using WIM data for research purposes (TRB 2014, 9). The USDOT report outlines the complex series of computations and assumptions leading to the baseline estimates of VMT by truck category (Modal Shift, 236–241). However, no tests of the estimates’ validity or accuracy are reported. Recommendations The research and data collection programs that would make the greatest contribution to future evaluations of truck size and weight regulations would be aimed at monitoring the impact of existing truck traffic

21 (and changes in traffic) on highway performance and at supporting state and local highway agencies’ efforts to manage the impacts. USDOT should promote and support development and improvement of these information systems. Improvements in information systems would have the secondary benefits of supporting more credible estimates of the effects of proposed changes in truck regulations. To improve regulations, the most critical information needs are for better understanding of how truck traffic characteristics (and changes in truck traffic over time) affect (a) bridge condition and highway agency costs related to bridges and (b) highway crash and casualty risks. To project infrastructure and safety impacts of regulatory changes, information on the determinants of vehicle and mode choice is necessary. USDOT and state research on truck regulatory policy should aim at evaluating the full range of methods for mitigating costs of truck traffic—not only size and weight limits but also changes in vehicle design; changes in bridge design, condition monitoring, and inspection practices; enforcement of dimensional and safety regulations; and design and management of truck permit and fee systems. The goal of research should be development of comprehensive strategies for improving the performance of highway freight transportation. Size and weight limits alone provide only weak leverage for improving performance. Future truck size and weight studies should be organized as evaluations of comprehensive policy options rather than evaluations of alternative truck configurations. Research aimed primarily at improved methods for predicting the impact of proposed changes in limits may have value for guiding policy. However, confidence that regulations are effective can come only from the ability to observe the impact of existing rules and the actual consequences of changes. References Abbreviations TRB Transportation Research Board USDOT U.S. Department of Transportation TRB. 1986. Special Report 211: Twin Trailer Trucks: Effects on Highways and Highway Safety. National

22 Research Council, Washington, D.C. TRB. 1990a. Special Report 225: Truck Weight Limits: Issues and Options. National Research Council, Washington, D.C. TRB. 1990b. Special Report 227: New Trucks for Greater Productivity and Less Road Wear: An Evaluation of the Turner Proposal. National Research Council, Washington, D.C. TRB. 2002. Special Report 267: Regulation of Weights, Lengths, and Widths of Commercial Motor Vehicles. National Academies, Washington, D.C. TRB. 2014. Review of U.S. Department of Transportation Truck Size and Weight Study: First Report: Review of Desk Scans. March 31. http://onlinepubs.trb.org/onlinepubs/sr/TS&WDeskScans.pdf. USDOT. 1981. An Investigation of Truck Size and Weight Limits: Final Report. Aug. USDOT. 2000. The U.S. Department of Transportation’s Comprehensive Truck Size and Weight Study: Volume III: Scenario Analysis. Aug.